audit: the alcohol use disorders identification test

Thomas F. BaborJohn C. Higgins-Biddle

John B. SaundersMaristela G. Monteiro

World Health Organization

A U D I T

The Alcohol Use Disorders Identification Test

Guidelines for Use in Primary Care

S e c o n d E d i t i o n

WHO/MSD/MSB/01.6aOriginal: English

Distribution: General

Department of Mental Health and Substance Dependence

Department of Mental Health and Substance Dependence

Thomas F. BaborJohn C. Higgins-Biddle

John B. SaundersMaristela G. Monteiro

World Health Organization

A U D I T

The Alcohol Use Disorders Identification Test

Guidelines for Use in Primary Care

S e c o n d E d i t i o n

WHO/MSD/MSB/01.6aOriginal: English

Distribution: General

2 I AUDIT I THE ALCOHOL USE DISORDERS IDENTIFICATION TEST

AAbbssttrraacctt

This manual introduces the AUDIT, the Alcohol Use Disorders Identification Test, and describes how touse it to identify persons with hazardous and harmful patterns of alcohol consumption. The AUDIT wasdeveloped by the World Health Organization (WHO) as a simple method of screening for excessive drinkingand to assist in brief assessment. It can help in identifying excessive drinking as the cause of the presentingillness. It also provides a framework for intervention to help hazardous and harmful drinkers reduce or ceasealcohol consumption and thereby avoid the harmful consequences of their drinking. The first edition of thismanual was published in 1989 (Document No. WHO/MNH/DAT/89.4) and was subsequently updated in1992 (WHO/PSA/92.4). Since that time it has enjoyed widespread use by both health workers and alcoholresearchers. With the growing use of alcohol screening and the international popularity of the AUDIT,there was a need to revise the manual to take into account advances in research and clinical experience.

This manual is written primarily for health care practitioners, but other professionals who encounter personswith alcohol-related problems may also find it useful. It is designed to be used in conjunction with acompanion document that provides complementary information about early intervention procedures, entitled“Brief Intervention for Hazardous and Harmful Drinking: A Manual for Use in Primary Care”. Togetherthese manuals describe a comprehensive approach to screening and brief intervention for alcohol-relatedproblems in primary health care.

AAcckknnoowwlleeddggeemmeennttss

The revision and finalisation of this document were coordinated by Maristela Monteiro with technicalassistance from Vladimir Poznyak from the WHO Department of Mental Health and Substance Dependence,and Deborah Talamini, University of Connecticut. Financial support for this publication was provided bythe Ministry of Health and Welfare of Japan.

© World Health Organization 2001

This document is not a formal publication of the World Health Organization (WHO), and all rights arereserved by the Organization. The document may, however, be freely reviewed, abstracted, reproduced,and translated, in part or in whole but not for sale or for use in conjunction with commercial purposes.Inquiries should be addressed to the Department of Mental Health and Substance Dependence, WorldHealth Organization, CH-1211 Geneva 27, Switzerland, which will be glad to provide the latest informationon any changes made to the text, plans for new editions and the reprints, regional adaptations and trans-lations that are already available.

Authors alone are responsible for views expressed in this document, which are not necessarily those ofthe World Health Organization.

TABLE OF CONTENTS I 3

Table of Contents

Purpose of this Manual

Why Screen for Alcohol Use?

The Context of Alcohol Screening

Development and Validation of the AUDIT

Administration Guidelines

Scoring and Interpretation

How to Help Patients

Programme Implementation

Appendix

A. Research Guidelines for the AUDIT

B. Suggested Format for AUDIT Self-Report Questionnaire

C. Translation and Adaptation to Specific Languages,Cultures and Standards

D. Clinical Screening Procedures

E. Training Materials for AUDIT

References

4

5

8

10

14

19

21

25

28

30

32

33

34

35

This manual introduces the AUDIT, theAlcohol Use Disorders Identification

Test, and describes how to use it to identifypersons with hazardous and harmful pat-terns of alcohol consumption. The AUDITwas developed by the World HealthOrganization (WHO) as a simple methodof screening for excessive drinking and toassist in brief assessment.1,2 It can helpidentify excessive drinking as the cause of the presenting illness. It provides a framework for intervention to help riskydrinkers reduce or cease alcohol con-sumption and thereby avoid the harmfulconsequences of their drinking. The AUDITalso helps to identify alcohol dependenceand some specific consequences of harm-ful drinking. It is particularly designed forhealth care practitioners and a range ofhealth settings, but with suitable instruc-tions it can be self-administered or usedby non-health professionals.

To this end, the manual will describe:

■ Reasons to ask about alcohol consumption

■ The context of alcohol screening

■ Development and validation of the AUDIT

■ The AUDIT questions and how to use them

■ Scoring and interpretation

■ How to conduct a clinical screeningexamination

■ How to help patients who screen positive

■ How to implement a screening programme

Purpose of this Manual

The appendices to this manual containadditional information useful to practi-tioners and researchers. Further researchon the reliability, validity, and implemen-tation of screening with the AUDIT issuggested using guidelines outlined inAppendix A. Appendix B contains anexample of the AUDIT in a self-reportquestionnaire format. Appendix C pro-vides guidelines for the translation andadaptation of the AUDIT. Appendix Ddescribes clinical screening proceduresusing a physical exam, laboratory testsand medical history data. Appendix E listsinformation about available trainingmaterials.


There are many forms of excessivedrinking that cause substantial risk or

harm to the individual. They include highlevel drinking each day, repeatedepisodes of drinking to intoxication,drinking that is actually causing physicalor mental harm, and drinking that hasresulted in the person becoming depen-dent or addicted to alcohol. Excessivedrinking causes illness and distress to thedrinker and his or her family and friends.It is a major cause of breakdown in rela-tionships, trauma, hospitalization, pro-longed disability and early death.Alcohol-related problems represent animmense economic loss to many commu-nities around the world.

AUDIT was developed to screen forexcessive drinking and in particular tohelp practitioners identify people whowould benefit from reducing or ceasingdrinking. The majority of excessivedrinkers are undiagnosed. Often theypresent with symptoms or problems thatwould not normally be linked to theirdrinking. The AUDIT will help the practi-tioner identify whether the person hashazardous (or risky) drinking, harmfuldrinking, or alcohol dependence.

Hazardous drinking3 is a pattern of alco-hol consumption that increases the riskof harmful consequences for the user orothers. Hazardous drinking patterns areof public health significance despite theabsence of any current disorder in theindividual user.

Harmful use refers to alcohol consump-tion that results in consequences to phys-ical and mental health. Some would alsoconsider social consequences among theharms caused by alcohol3, 4.

Alcohol dependence is a cluster ofbehavioural, cognitive, and physiologicalphenomena that may develop afterrepeated alcohol use4. Typically, thesephenomena include a strong desire toconsume alcohol, impaired control overits use, persistent drinking despite harm-ful consequences, a higher priority givento drinking than to other activities andobligations, increased alcohol tolerance,and a physical withdrawal reaction whenalcohol use is discontinued.

Alcohol is implicated in a wide variety ofdiseases, disorders, and injuries, as well asmany social and legal problems5,6,7. It is amajor cause of cancer of the mouth,esophagus, and larynx. Liver cirrhosis andpancreatitis often result from long-term,excessive consumption. Alcohol causesharm to fetuses in women who are preg-nant. Moreover, much more commonmedical conditions, such as hypertension,gastritis, diabetes, and some forms ofstroke are likely to be aggravated even byoccasional and short-term alcohol con-sumption, as are mental disorders such asdepression. Automobile and pedestrianinjuries, falls, and work-related harm fre-quently result from excessive alcohol con-sumption. The risks related to alcohol arelinked to the pattern of drinking and theamount of consumption5. While personswith alcohol

WHY SCREEN FOR ALCOHOL USE? I 5

Why Screen for Alcohol Use?

dependence are most likely to incur highlevels of harm, the bulk of harm associat-ed with alcohol occurs among people whoare not dependent, if only because thereare so many of them8. Therefore, theidentification of drinkers with varioustypes and degrees of at-risk alcohol con-sumption has great potential to reduce alltypes of alcohol-related harm.

Figure 1 illustrates the large variety ofhealth problems associated with alcoholuse. Although many of these medicalconsequences tend to be concentrated inpersons with severe alcohol dependence,even the use of alcohol in the range of20-40 grams of absolute alcohol per dayis a risk factor for accidents, injuries, andmany social problems5, 6.

Many factors contribute to the develop-ment of alcohol-related problems.Ignorance of drinking limits and of therisks associated with excessive alcoholconsumption are major factors. Socialand environmental influences, such ascustoms and attitudes that favor heavydrinking, also play important roles. Ofutmost importance for screening, however,is the fact that people who are notdependent on alcohol may stop orreduce their alcohol consumption withappropriate assistance and effort. Oncedependence has developed, cessation of alcohol consumption is more difficultand often requires specialized treatment.Although not all hazardous drinkersbecome dependent, no one developsalcohol dependence without havingengaged for some time

in hazardous alcohol use. Given thesefactors, the need for screening becomesapparent.

Screening for alcohol consumptionamong patients in primary care carriesmany potential benefits. It provides anopportunity to educate patients aboutlow-risk consumption levels and the risksof excessive alcohol use. Informationabout the amount and frequency of alco-hol consumption may inform the diagno-sis of the patient’s presenting condition,and it may alert clinicians to the need toadvise patients whose alcohol consump-tion might adversely affect their use ofmedications and other aspects of theirtreatment. Screening also offers theopportunity for practitioners to take pre-ventative measures that have proveneffective in reducing alcohol-related risks.


WHY SCREEN FOR ALCOHOL USE? I 7

Figure 1

Effects of High-Risk Drinking

Numb, tingling toes.Painful nerves.

Impaired sensationleading to falls.

Inflammation of the pancreas.

Vitamin deficiency. Bleeding.Severe inflammation

of the stomach. Vomiting.Diarrhea. Malnutrition.

Cancer of throat and mouth .

Premature aging. Drinker's nose.

Weakness of heart muscle.Heart failure. Anemia.

Impaired blood clotting.Breast cancer.

In men: Impaired sexual performance.

In women:Risk of giving birth to deformed,

retarded babies or low birthweight babies.

Aggressive,irrational behaviour.Arguments. Violence.Depression. Nervousness.

Frequent colds. Reducedresistance to infection.Increased risk of pneumonia.

Alcohol dependence.Memory loss.

Ulcer.

Liver damage.

Trembling hands.Tingling fingers.Numbness. Painful nerves.

High-risk drinking may lead to social, legal, medical, domestic, job and financialproblems. It may also cut your lifespan and lead to accidents and death from drunk-en driving.

The Context of Alcohol Screening


While this manual focuses on using theAUDIT to screen for alcohol con-

sumption and related risks in primary caremedical settings, the AUDIT can be effec-tively applied in many other contexts aswell. In many cases procedures havealready been developed and used in thesesettings. Box 1 summarizes informationabout the settings, screening personnel,and target groups considered appropriatefor a screening programme using the AUDIT.Murray9 has argued that screening mightbe conducted profitably with:

■ general hospital patients, especially thosewith disorders known to be associatedwith alcohol dependence (e.g., pancre-atitis, cirrhosis, gastritis, tuberculosis,neurological disorders, cardiomyopathy);

■ persons who are depressed or whoattempt suicide;

■ other psychiatric patients;

■ patients attending casualty and emer-gency services;

■ patients attending general practitioners;

■ vagrants;

■ prisoners; and

■ those cited for legal offences connectedwith drinking (e.g., driving while intoxi-cated, public intoxication).

To these should be added groups consid-ered by a WHO Expert Committee7 to beat high risk of developing alcohol-relatedproblems: middle-aged males, adolescents,migrant workers, and certain occupation-al groups (such as business executives,entertainers, sex workers, publicans, andseamen). The nature of the risk differs byage, gender, drinking context, and drinkingpattern, with sociocultural factors playingan important role in the definition andexpression of alcohol-related problems6.

THE CONTEXT OF ALCOHOL SCREENING I 9

Box 1

Personnel, Settings and Groups Considered Appropriate for aScreening Programme Using the AUDIT

Setting Target Group Screening Personnel

Primary care clinic Medical patients Nurse, social worker

Emergency room Accident victims, Physician, nurse, or staffIntoxicated patients,trauma victims

Physician’s Room Medical patients General practitioner,Surgery family physician or staff

General Hospital wards Patients with Internist, staffOut-patient clinic hypertension, heart

disease, gatrointestinalor neurological disorders

Psychiatric hospital Psychiatric patients, Psychiatrist, staffparticularly thosewho are suicidal

Court, jail, prison DWI offenders Officers, Counsellorsviolent criminals

Other health-related Persons demonstrating Health and human facilities impaired social or service workers

occupational functioning(e.g. marital discord,child neglect, etc.)

Military Services Enlisted men and officers Medics

Work place Workers, especially those Employee assistance staffEmployee assistance having problems withProgramme productivity, absenteeism

or accidents


Development and Validation of the AUDIT

The AUDIT was developed and evaluat-ed over a period of two decades, and

it has been found to provide an accuratemeasure of risk across gender, age, andcultures1, 2,10. Box 2 describes the conceptu-al domains and item content of the AUDIT,which consists of 10 questions aboutrecent alcohol use, alcohol dependencesymptoms, and alcohol-related problems.As the first screening test designed specif-ically for use in primary care settings, theAUDIT has the following advantages:

■ Cross-national standardization: theAUDIT was validated on primary healthcare patients in six countries1,2. It is theonly screening test specifically designedfor international use;

■ Identifies hazardous and harmful alco-hol use, as well as possible dependence;

■ Brief, rapid, and flexible;

■ Designed for primary health care workers;

■ Consistent with ICD-10 definitions of alco-hol dependence and harmful alcohol use3,4;

■ Focuses on recent alcohol use.

In 1982 the World Health Organizationasked an international group of investiga-tors to develop a simple screening instru-ment2. Its purpose was to identify personswith early alcohol problems using proce-dures that were suitable for health systemsin both developing and developed countries.The investigators reviewed a variety ofself-report, laboratory, and clinical proce-dures that had been used for this purposein different countries. They then initiated across-national study to select the best fea-tures of these various national approachesto screening1.

This comparative field study was conductedin six countries (Norway, Australia, Kenya,Bulgaria, Mexico, and the United Statesof America).

The method consisted of selecting itemsthat best distinguished low-risk drinkersfrom those with harmful drinking. Unlikeprevious screening tests, the new instrumentwas intended for the early identification ofhazardous and harmful drinking as well asalcohol dependence (alcoholism). Nearly2000 patients were recruited from a varietyof health care facilities, including specializedalcohol treatment centers. Sixty-four percentwere current drinkers, 25% of whom werediagnosed as alcohol dependent.

Participants were given a physical exami-nation, including a blood test for standardblood markers of alcoholism, as well as anextensive interview assessing demographiccharacteristics, medical history, healthcomplaints, use of alcohol and drugs, psy-chological reactions to alcohol, problemsassociated with drinking, and family histo-ry of alcohol problems. Items were select-ed for the AUDIT from this pool of ques-tions primarily on the basis of correlationswith daily alcohol intake, frequency ofconsuming six or more drinks per drinkingepisode, and their ability to discriminatehazardous and harmful drinkers. Items werealso chosen on the basis of face validity,clinical relevance, and coverage of relevantconceptual domains (i.e., alcohol use, alco-hol dependence, and adverse consequencesof drinking). Finally, special attention in itemselection was given to gender appropriate-ness and cross-national generalizability.

Sensitivities and specificities of the select-ed test items were computed for multiplecriteria (i.e., average daily alcohol consump-tion, recurrent intoxication, presence of atleast one dependence symptom, diagnosisof alcohol abuse or dependence, and self-perception of a drinking problem). Variouscut-off points in total scores were consid-ered to identify the value with optimalsensitivity (percentage of positive casesthat the test correctly identified) andspecificity (percentage of negative casesthat the test correctly identified) to distin-guish hazardous and harmful alcohol use.In addition, validity was also computedagainst a composite diagnosis of harmfuluse and dependence. In the test develop-

ment samples1, a cut-off value of 8 pointsyielded sensitivities for the AUDIT for vari-ous indices of problematic drinking thatwere generally in the mid 0.90’s.Specificities across countries and acrosscriteria averaged in the 0.80’s.

The AUDIT differs from other self-reportscreening tests in that it was based ondata collected from a large multinationalsample, used an explicit conceptual-statistical rationale for item selection,emphasizes identification of hazardousdrinking rather than long-term dependenceand adverse drinking consequences, andfocuses primarily on symptoms occurringduring the recent past rather than “ever.”

DEVELOPMENT AND VALIDATION OF THE AUDIT I 11

Box 2

Domains and Item Content of the AUDIT

Domains Question Item ContentNumber

Hazardous 1 Frequency of drinkingAlcohol 2 Typical quantity

Use 3 Frequency of heavy drinking

Dependence 4 Impaired control over drinkingSymptoms 5 Increased salience of drinking

6 Morning drinking

Harmful 7 Guilt after drinkingAlcohol 8 Blackouts

Use 9 Alcohol-related injuries10 Others concerned about

drinking

Once the AUDIT had been published, thedevelopers recommended additional vali-dation research. In response to this request,a large number of studies have been con-ducted to evaluate its validity and reliabil-ity in different clinical and communitysamples throughout the world10. At therecommended cut-off of 8, most studieshave found very favorable sensitivity andusually lower, but still acceptable, speci-ficity, for current ICD-10 alcohol use dis-orders10,11,12 as well as the risk of futureharm12. Nevertheless, improvements indetection have been achieved in somecases by lowering or raising the cut-offscore by one or two points, dependingon the population and the purpose ofthe screening programme11,12.

A variety of subpopulations have been stud-ied, including primary care patients 13, 14, 15,emergency room cases11, drug users16,the unemployed17, university students18,elderly hospital patients19, and persons oflow socio-economic status20. The AUDIThas been found to provide good discrimi-nation in a variety of settings where thesepopulations are encountered. A recentsystematic review21 of the literature hasconcluded that the AUDIT is the bestscreening instrument for the whole rangeof alcohol problems in primary care, ascompared to other questionnaires such asthe CAGE and the MAST.

Cultural appropriateness and cross-national applicability were important con-siderations in the development of theAUDIT1, 2. Research has been conductedin a wide variety of countries and

cultures11, 12, 13, 15, 19, 22, 23, 24, suggestingthat the AUDIT has fulfilled its promise asan international screening test.

Although evidence on women is somewhatlimited11, 12, 24, the AUDIT seems equallyappropriate for males and females. Theeffect of age has not been systematicallyanalyzed as a possible influence on theAUDIT, but one study19 found low sensi-tivity but high specificity in patients aboveage 65. The AUDIT has proven to beaccurate in detecting alcohol dependencein university students18.

In comparison to other screening tests,the AUDIT has been found to performequally well or at a higher degree of accu-racy10, 11, 25, 26 across a wide variety of cri-terion measures. Bohn, et al.27 found astrong correlation between the AUDITand the MAST (r=.88) for both males andfemales, and correlations of .47 and .46for males and females, respectively, on acovert content alcoholism screening test.A high correlation coefficient (.78) wasalso found between the AUDIT and theCAGE in ambulatory care patients26.AUDIT scores were found to correlate wellwith measures of drinking consequences,attitudes toward drinking, vulnerability toalcohol dependence, negative mood statesafter drinking, and reasons for drinking27.It appears that the total score on the AUDITreflects the extent of alcohol involvementalong a broad continuum of severity.

Two studies have considered the relationbetween AUDIT scores and future indica-tors of alcohol-related problems and more


global life functioning. In one study17, thelikelihood of remaining unemployed overa two year period was 1.6 times higherfor individuals with scores of 8 or moreon the AUDIT than for comparable per-sons with lower scores. In another study28,AUDIT scores of ambulatory care patientspredicted future occurrence of a physicaldisorder, as well as social problems relat-ed to drinking. AUDIT scores also predict-ed health care utilization and future riskof engaging in hazardous drinking28.

Several studies have reported on the reli-ability of the AUDIT18, 26, 29. The resultsindicate high internal consistency, suggest-ing that the AUDIT is measuring a singleconstruct in a reliable fashion. A test-retestreliability study29 indicated high reliability(r=.86) in a sample consisting of non-haz-ardous drinkers, cocaine abusers, andalcoholics. Another methodological studywas conducted in part to investigate theeffect of question ordering and wordingchanges on prevalence estimates andinternal consistency reliability22. Changesin question ordering and wording did notaffect the AUDIT scores, suggesting thatwithin limits, researchers can exercisesome flexibility in modifying the orderand wording of the AUDIT items.

With increasing evidence of the reliabilityand validity of the AUDIT, studies havebeen conducted using the test as aprevalence measure. Lapham, et al.23

used it to estimate prevalence of alcoholuse disorders in emergency rooms (ERs)of three regional hospitals in Thailand.

It was concluded that the ER is an idealsetting for implementing alcohol screen-ing with the AUDIT. Similarly, Piccinelli, etal.15 evaluated the AUDIT as a screeningtool for hazardous alcohol intake in prima-ry care clinics in Italy. AUDIT performedwell in identifying alcohol-related disordersas well as hazardous use. Ivis, et al.22

incorporated the AUDIT into a generalpopulation telephone survey in Ontario,Canada.

Since the AUDIT User’s Manual was firstpublished in 198930, the test has fulfilledmany of the expectations that inspired itsdevelopment. Its reliability and validity havebeen established in research conducted ina variety of settings and in many differentnations. It has been translated into manylanguages, including Turkish, Greek, Hindi,German, Dutch, Polish, Japanese, French,Portuguese, Spanish, Danish, Flemish,Bulgarian, Chinese, Italian, and Nigeriandialects. Training programmes have beendeveloped to facilitate its use by physiciansand other health care providers31, 32 (seeAppendix E). It has been used in primarycare research and in epidemiologicalstudies for the estimation of prevalencein the general population as well as spe-cific institutional groups (e.g., hospitalpatients, primary care patients). Despitethe high level of research activity on theAUDIT, further research is needed, espe-cially in the less developed countries.Appendix A provides guidelines for con-tinued research on the AUDIT.

DEVELOPMENT AND VALIDATION OF THE AUDIT I 13


Administration Guidelines

The AUDIT can be used in a variety ofways to assess patients’ alcohol use,

but programmes to implement it shouldfirst set guidelines that consider thepatient’s circumstances and capacities.Additionally, care must be taken to tellpatients why questions about alcohol useare being asked and to provide informa-tion they need to make appropriateresponses. A decision must be madewhether to administer the AUDIT orally oras a written, self-report questionnaire.Finally, consideration must be given tousing skip-outs to shorten the screeningfor greater efficiency. This section recom-mends guidelines on such issues ofadministration.

Considering the Patient

All patients should be screened for alco-hol use, preferably annually. The AUDITcan be administered separately or com-bined with other questions as part of ageneral health interview, a lifestyle ques-tionnaire, or medical history. If healthworkers screen only those they considermost likely to have a “drinking problem”,the majority of patients who drink exces-sively will be missed. However, it is impor-tant to consider the condition of thepatients when asking them to answerquestions about alcohol use. To increasethe patient’s receptivity to the questionsand the accuracy of responding, it isimportant that:

■ The interviewer (or presenter of the sur-vey) be friendly and non-threatening;

■ The patient is not intoxicated or in needof emergency care at the time;

■ The purpose of the screening be clearlystated in terms of its relevance to thepatient’s health status;

■ The information patients need tounderstand the questions and respondaccurately be provided; and

■ Assurance is given that the patient’sresponses will remain confidential.

Health workers should try to establishthese conditions before the AUDIT isgiven. When these conditions are not pre-sent or when a patient is resistant, theClinical Screening Procedures (discussed inAppendix D) may provide an alternativecourse of action.

Choose the best possible circumstance foradministering the AUDIT. For patientsrequiring emergency treatment or in greatpain, it is best to wait until their medicalcondition has stabilized and they havebecome accustomed to the health settingwhere administration of the AUDIT is totake place. Look for signs of alcohol ordrug intoxication. Patients who have alco-hol on their breath or who appear intoxi-cated may be unreliable respondents.Consider screening at a later time. If thisis not possible, make note of these find-ings on the patient's record.

When presented in a medical contextwith genuine concern for the patient’swell being, patients are almost alwaysopen and responsive to the AUDIT ques-tions. Moreover, most patients answer thequestions honestly. Even when excessive

ADMINISTRATION GUIDELINES I 15

drinkers underestimate their consump-tion, they often qualify on the AUDITscoring system as positive for alcohol risk.

Introducing the AUDIT

Whether the AUDIT is used as an oralinterview or a written questionnaire, it isrecommended that an explanation begiven to patients of the content of thequestions, the purpose for asking them,and the need for accurate answers. Thefollowing are illustrative introductions fororal delivery and written questionnaires:

“Now I am going to ask you some ques-tions about your use of alcoholic bever-ages during the past year. Because alco-hol use can affect many areas of health(and may interfere with certain medica-tions), it is important for us to know howmuch you usually drink and whether youhave experienced any problems with yourdrinking. Please try to be as honest andas accurate as you can be.”

“As part of our health service it is impor-tant to examine lifestyle issues likely toaffect the health of our patients. Thisinformation will assist in giving you thebest treatment and highest possible stan-dard of care. Therefore, we ask that youcomplete this questionnaire that asksabout your use of alcoholic beveragesduring the past year. Please answer asaccurately and honestly as possible. Yourhealth worker will discuss this issue withyou. All information will be treated instrict confidence.

This statement should be followed by adescription of the types of alcoholic bev-erages typically consumed in the countryor region where the patient lives (e.g., “Byalcoholic beverages we mean your use ofwine, beer, vodka, sherry, etc.”) If neces-sary, include a description of beveragesthat may not be considered alcoholic,(e.g. cider, low alcohol beer, etc.). Withpatients whose alcohol consumption isprohibited by law, culture, or religion(e.g., youths, observant Muslims), acknowl-edgment of such prohibition and encour-agement of candor may be needed. Forexample, “I understand others may thinkyou should not drink alcohol at all, but itis important in assessing your health toknow what you actually do.”

Patient instructions should also clarify themeaning of a standard drink. Questions 2 and 3 of AUDIT ask about “drinks con-sumed”. The meaning of this word differsfrom one nation and culture to another.It is important therefore to mention themost common alcoholic beverages likelyto be consumed and how much of eachconstitutes a drink (approximately 10 gramsof pure ethanol). For example, one bottleof beer (330 ml at 5% ethanol), a glassof wine (140 ml at 12% ethanol), and ashot of spirits (40 ml at 40% ethanol)represent a standard drink of about 13 gof ethanol. Since the types and amountsof alcoholic drinks will vary according toculture and custom, the alcohol contentof typical servings of beer, wine and spiritsmust be determined to adapt the AUDITto particular settings. See Appendix C.

Oral Administration vs.Self-report Questionnaire

The AUDIT may be administered either asan oral interview or as a self-report ques-tionnaire. Each method carries its ownadvantages and disadvantages that mustbe weighed in light of time and cost con-straints. The relative merits of using theAUDIT as an interview vs. the self-reportquestionnaire are summarized in Box 3.The cognitive capacities (literacy, forgetful-ness) and level of cooperation (defensive-ness) of the patient should be considered.If the expectation is that primary careproviders will manage all the care thatpatients will receive for their alcohol prob-lems, an interview may have advantages.However, if the provider’s responsibilitywill be limited to offering brief advice topatients who screen positive and referringmore severe cases to other services, thequestionnaire method may be preferable.

Whatever decision is made, it must be con-sistent with implementation plans to estab-lish a comprehensive screening programme.

The AUDIT questions and responses arepresented in Box 4 in a format suggestedfor an oral interview. Appendix B gives anexample of the self-report questionnaire.Adaptation should be made to needs ofthe particular screening programme as wellas the alcoholic beverages most commonlyconsumed in that society. Appendix C pro-vides guidelines for translation and adapta-tion to national and local conditions.

If the AUDIT is administered as an interview,it is important to read the questions aswritten and in the order indicated. By fol-lowing the exact wording, better compa-rability will be obtained between yourresults and those obtained by other inter-viewers. Most of the questions in theAUDIT are phrased in terms of “how


Box 3

Advantages of Different Approaches to AUDIT Administration

Questionnaire Interview

Takes less time Allows clarification of ambiguous answers

Easy to administer Can be administered to patients with poor reading skills

Suitable for computer administration and scoring

May produce more accurate answers Allows seamless feedback to patient and initiation of brief advice

ADMINISTRATION GUIDELINES I 17

Box 4

The Alcohol Use Disorders Identification Test: Interview VersionRead questions as written. Record answers carefully. Begin the AUDIT by saying “Now I am going to ask you some questions about your use of alcoholic beverages during this past year.” Explain what is meant by “alcoholic beverages” by using local examples of beer, wine, vodka, etc. Code answers in terms of “standard drinks”. Place the correct answer number in the box at the right.

6. How often during the last year have you neededa first drink in the morning to get yourself goingafter a heavy drinking session?

(0) Never(1) Less than monthly(2) Monthly(3) Weekly(4) Daily or almost daily

1. How often do you have a drink containing alco-hol?

(0) Never [Skip to Qs 9-10](1) Monthly or less(2) 2 to 4 times a month(3) 2 to 3 times a week(4) 4 or more times a week

7. How often during the last year have you had afeeling of guilt or remorse after drinking?


4. How often during the last year have you foundthat you were not able to stop drinking once youhad started?


9. Have you or someone else been injured as aresult of your drinking?

(0) No(2) Yes, but not in the last year(4) Yes, during the last year

5. How often during the last year have you failed todo what was normally expected from youbecause of drinking?


10. Has a relative or friend or a doctor or anotherhealth worker been concerned about your drink-ing or suggested you cut down?

(0) No(2) Yes, but not in the last year(4) Yes, during the last year

2. How many drinks containing alcohol do you haveon a typical day when you are drinking?

(0) 1 or 2(1) 3 or 4(2) 5 or 6(3) 7, 8, or 9(4) 10 or more

8. How often during the last year have you beenunable to remember what happened the nightbefore because you had been drinking?


3. How often do you have six or more drinks on oneoccasion?


Skip to Questions 9 and 10 if Total Scorefor Questions 2 and 3 = 0

Record total of specific items hereIf total is greater than recommended cut-off, consult User’s Manual.

often” symptoms occur. Provide thepatient with the response categories givenfor each question (for example, “Never,”“Several times a month,” “Daily”). When aresponse option has been chosen, it isuseful to probe during the initial ques-tions to be sure that the patient hasselected the most accurate response (forexample, “You say you drink several timesa week. Is this just on weekends or doyou drink more or less every day?”).

If responses are ambiguous or evasive,continue asking for clarification by repeat-ing the question and the response options,asking the patient to choose the bestone. At times answers are difficult torecord because the patient may not drinkon a regular basis. For example, if thepatient was drinking excessively duringthe month before an accident, but notprior to that time, then it will be difficultto characterize the “typical” drinkingsought by the question. In these cases itis best to record the amount of drinkingand related symptoms for the heaviestdrinking period in the past year, makingnote of the fact that this may be atypicalor transitory for that individual.

Record answers carefully, making note ofany special circumstances, additional infor-mation, and clinical observations. Oftenpatients will provide the interviewer withuseful comments about their drinking thatcan be valuable in the interpretation of theAUDIT total score.

Administering the AUDIT as a written ques-tionnaire or by computer eliminates manyof the uncertainties of patient responses byallowing only specific choices.

However, it eliminates the informationobtained from the interview format.Moreover, it presumes literacy and abilityof the patient to perform the requiredactions. It may also require less time onthe part of health workers, if patients cancomplete the process alone. With time ata premium for both health workers andpatients, ways of shortening the screeningprocess merit consideration.

Shortening the Screening ProcessAdministered either orally or as a question-naire, the AUDIT can usually be completedin two to four minutes and scored in afew seconds. However, for many patientsit is unnecessary to administer the completeAUDIT because they drink infrequently,moderately, or abstain entirely from alco-hol. The interview version of the AUDIT(Box 4) provides two opportunities to skipquestions for such patients. If the patientanswers in response to Question 1 that nodrinking has occurred during the last year,the interviewer may skip to Questions 9-10,responses to which may indicate past prob-lems with alcohol. Patients who score pointson these questions may be considered at riskif they begin to drink again, and should beadvised to avoid alcohol. It is recommendedthat this skip out instruction only be usedwith the interview or computer-assistedformats of the AUDIT.

A second opportunity to shorten AUDITscreening occurs after Question 3 hasbeen answered. If the patient scored 0 on Questions 2 and 3, the interviewermay skip to Questions 9-10 because thepatient’s drinking has not exceeded thelow risk drinking limits.


The AUDIT is easy to score. Each of thequestions has a set of responses to

choose from, and each response has ascore ranging from 0 to 4. In the interviewformat (Box 4) the interviewer enters thescore (the number within parentheses)corresponding to the patient’s responseinto the box beside each question. In the self-report questionnaire format(Appendix B), the number in the columnof each response checked by the patientshould be entered by the scorer in theextreme right-hand column. All the responsescores should then be added and recordedin the box labeled “Total”.

Total scores of 8 or more are recom-mended as indicators of hazardous andharmful alcohol use, as well as possiblealcohol dependence. (A cut-off score of10 will provide greater specificity but atthe expense of sensitivity.) Since theeffects of alcohol vary with average bodyweight and differences in metabolism,establishing the cut off point for allwomen and men over age 65 one pointlower at a score of 7 will increase sensi-tivity for these population groups.Selection of the cut-off point should beinfluenced by national and cultural stan-dards and by clinician judgment, whichalso determine recommended maximumconsumption allowances. Technicallyspeaking, higher scores simply indicategreater likelihood of hazardous andharmful drinking. However, such scoresmay also reflect greater severity of alcoholproblems and dependence, as well as agreater need for more intensive treatment.

More detailed interpretation of a patient’stotal score may be obtained by determin-ing on which questions points werescored. In general, a score of 1 or moreon Question 2 or Question 3 indicatesconsumption at a hazardous level. Pointsscored above 0 on questions 4-6 (espe-cially weekly or daily symptoms) imply thepresence or incipience of alcohol depen-dence. Points scored on questions 7-10indicate that alcohol-related harm isalready being experienced. The totalscore, consumption level, signs of depen-dence, and present harm all should playa role in determining how to manage apatient. The final two questions shouldalso be reviewed to determine whetherpatients give evidence of a past problem(i.e., “yes, but not in the past year”).Even in the absence of current hazardousdrinking, positive responses on theseitems should be used to discuss the needfor vigilance by the patient.

In most cases the total AUDIT score willreflect the patient’s level of risk related toalcohol. In general health care settingsand in community surveys, most patientswill score under the cut-offs and may beconsidered to have low risk of alcohol-related problems. A smaller, but still sig-nificant, portion of the population is like-ly to score above the cut-offs but recordmost of their points on the first threequestions. A much smaller proportioncan be expected to score very high, withpoints recorded on the dependence-relat-ed questions as well as exhibiting alco-hol-related problems. As yet there hasbeen insufficient research to establish

SCORING AND INTERPRETATION I 19

Scoring and Interpretation

precisely a cut-off point to distinguishhazardous and harmful drinkers (whowould benefit from a brief intervention)from alcohol dependent drinkers (whoshould be referred for diagnostic evalua-tion and more intensive treatment). Thisis an important question because screen-ing programmes designed to identifycases of alcohol dependence are likely tofind a large number of hazardous andharmful drinkers if the cut-off of 8 isused. These patients need to be man-aged with less intensive interventions. Ingeneral, the higher the total score on theAUDIT, the greater the sensitivity in find-ing persons with alcohol dependence.

Based on experience gained in a study oftreatment matching with persons whohad a wide range of alcohol problemseverity, AUDIT scores were comparedwith diagnostic data reflecting low, medi-um and high degrees of alcohol depen-dence. It was found that AUDIT scores inthe range of 8-15 represented a mediumlevel of alcohol problems whereas scoresof 16 and above represented a high levelof alcohol problems33. On the basis ofexperience gained from the use of theAUDIT in this and other research, it issuggested that the following interpreta-tion be given to AUDIT scores:

■ Scores between 8 and 15 are mostappropriate for simple advice focusedon the reduction of hazardous drinking.

■ Scores between 16 and 19 suggestbrief counseling and continued moni-toring.

■ AUDIT scores of 20 or above clearlywarrant further diagnostic evaluationfor alcohol dependence.

In the absence of better research theseguidelines should be considered tenta-tive, subject to clinical judgment thattakes into account the patient’s medicalcondition, family history of alcohol prob-lems and perceived honesty in respond-ing to the AUDIT questions.

While use of the 10-question AUDITquestionnaire will be sufficient for thevast majority of patients, special circum-stances may require a clinical screeningprocedure. For example, a patient may beresistant, uncooperative, or unable torespond to the AUDIT questions. If fur-ther confirmation of possible dependenceis warranted, a physical examination pro-cedure and laboratory tests may be used,as described in Appendix D.


Using the AUDIT to screen patients is onlythe first step in a process of helping

reduce alcohol-related problems and risks.

Health care workers must decide whatservices they can provide to patients whoscore positive. Once a positive case hasbeen identified, the next step is to providean appropriate intervention that meets theneeds of each patient. Typically, alcoholscreening has been used primarily to find“cases” of alcohol dependence, who arethen referred to specialized treatment. Inrecent years, however, advances in screen-ing procedures have made it possible toscreen for risk factors, such as hazardousdrinking and harmful alcohol use. Usingthe AUDIT Total Score, there is a simple wayto provide each patient with an appropri-ate intervention, based on the level of risk.

While this discussion will focus on helpingthose patients who score positive on theAUDIT, sound preventative practice alsocalls for reporting screening results tothose who score negative. These patientsshould be reminded about the benefits oflow risk drinking or abstinence and toldnot to drink in certain circumstances,such as those mentioned in Box 5.

Four levels of risk are shown in Box 6.Zone I refers to low risk drinking or absti-nence. The second level, Zone II, consistsof alcohol use in excess of low-risk guide-lines5, and is generally indicated when theAUDIT score is between 8 and 15. A briefintervention using simple advice and patienteducation materials is the most appropriatecourse of action for these patients. The

third level, Zone III, is suggested by AUDITscores in the range of 16 to 19. Harmfuland hazardous drinking can be managedby a combination of simple advice, briefcounseling and continued monitoring,with further diagnostic evaluation indicatedif the patient fails to respond or is suspectedof possible alcohol dependence. The fourthrisk level is suggested by AUDIT scores inexcess of 20. These patients should bereferred to a specialist for diagnostic evalu-ation and possible treatment for alcoholdependence. If these services are not avail-able, these patients can be managed inprimary care, especially when mutual helporganizations are able to provide commu-nity-based support. Using a stepped-careapproach, patients can be managed first atthe lowest level of intervention suggestedby their AUDIT score. If they do not respondto the initial intervention, they should bereferred to the next level of care.

HOW TO HELP PATIENTS I 21

How to Help Patients

Box 5

Advise Patients not to Drink

■ When operating a vehicle ormachinery

■ When pregnant or consideringpregnancy

■ If a contraindicated medical condition is present

■ After using certain medications,such as sedatives, analgesics, and selected antihypertensives

Brief interventions for hazardous andharmful drinking constitute a variety ofactivities characterized by their low inten-sity and short duration. They range from5 minutes of simple advice about how toreduce hazardous drinking to several ses-sions of brief counseling to address morecomplicated conditions36. Intended toprovide early intervention, before or soonafter the onset of alcohol-related problems,brief interventions consist of feedback ofscreening data designed to increase moti-vation to change drinking behaviour, aswell as simple advice, health education,skill building, and practical suggestions.Over the last 20 years procedures havebeen developed that primary care practi-tioners can readily learn and practice toaddress hazardous and harmful drinking.These procedures are summarized in Box 7.

A number of randomized controlled trialshave evaluated the efficacy of this approach,showing consistently positive benefits for


Box 7

Elements of BriefInterventions

■ Present screening results

■ Identify risks and discuss conse-quences

■ Provide medical advice

■ Solicit patient commitment

■ Identify goal—reduced drinking orabstinence

■ Give advice and encouragement

Box 6

Risk Level Intervention AUDIT score*

Zone I Alcohol Education 0-7

Zone II Simple Advice 8-15

Zone III Simple Advice plus Brief Counseling and Continued Monitoring 16-19

Zone IV Referral to Specialist for Diagnostic 20-40Evaluation and Treatment

*The AUDIT cut-off score may vary slightly depending on the country’s drinking patterns, the alcohol content ofstandard drinks, and the nature of the screening program. Clinical judgment should be exercised in cases wherethe patient’s score is not consistent with other evidence, or if the patient has a prior history of alcohol dependence.It may also be instructive to review the patient’s responses to individual questions dealing with dependence symp-toms (Questions 4, 5 and 6) and alcohol-related problems (Questions 9 and 10). Provide the next highest level ofintervention to patients who score 2 or more on Questions 4, 5 and 6, or 4 on Questions 9 or 10.

patients who are not dependent on alco-hol36, 37, 38. A companion WHO manual,Brief Intervention for Hazardous andHarmful Drinking: A Manual for Use inPrimary Care, provides more informationon this approach.

Referral to alcohol specialty care is commonamong those primary care practitionerswho do not have competency in treatingalcohol use disorders and where specialtycare is available. Consideration must begiven to the willingness of patients toaccept referral and treatment. Manypatients underestimate the risks associat-ed with drinking; others may not be pre-pared to admit and address their depen-dence. A brief intervention, adapted tothe purpose of initiating a referral usingdata from a clinical examination andblood tests, may help to address patientresistance. Follow-up with the patientand the specialty provider may alsoassure that the referral is accepted andtreatment is received.

Diagnosis is a necessary step followinghigh positive scoring on the AUDIT, sincethe instrument does not provide suffi-cient basis for establishing a manage-ment or treatment plan. While personsassociated with the screening programmeshould have a basic familiarity with thecriteria for alcohol dependence, a quali-fied professional who is trained in thediagnosis of alcohol use disorders4 shouldconduct this assessment. The bestmethod of establishing a diagnosis isthrough the use of a standardized, struc-tured, psychiatric interview, such as the

CIDI39 or the SCAN40. The alcohol sectionsof these interviews require 5 to 10 minutesto complete.

The Tenth revision of the InternationalClassification of Diseases (ICD-10)4 pro-vides detailed guidelines for the diagnosisof acute alcohol intoxication, harmful use,alcohol dependence syndrome, withdrawalstate, and related medical and neuropsy-chiatric conditions. The ICD-10 criteria forthe alcohol dependence syndrome aredescribed in Box 8.

Detoxification may be necessary for somepatients. Special attention should be paidto patients whose AUDIT responses indi-cate daily consumption of large amountsof alcohol and/or positive responses toquestions indicative of possible depen-dence (questions 4-6). Enquiry should bemade as to how long a patient has gonesince having an alcohol-free day and anyprior experience of withdrawal symp-toms. This information, a physical exami-nation, and laboratory tests (see ClinicalScreening Procedures, Appendix D) mayinform a judgment of whether to recom-mend detoxification. Detoxificationshould be provided for patients likely toexperience moderate to severe withdraw-al not only to minimize symptoms, butalso to prevent or manage seizures ordelirium, and to facilitate acceptance oftherapy to address dependence. Whileinpatient detoxification may be necessaryin a small number of severe cases, ambu-latory or home detoxification can be usedsuccessfully with the majority of lesssevere cases.

HOW TO HELP PATIENTS I 23

Medical management or treatment ofalcohol dependence has been describedin previous WHO publications41. A varietyof treatments for alcohol dependencehave been developed and foundeffective42. Significant advances havebeen made in pharmacotherapy, familyand social support therapy, relapse pre-vention, and behaviour-oriented skillstraining interventions.

Because the diagnosis and treatment ofalcohol dependence have developed as aspecialty within the mainstream of med-ical care, in most countries primary carepractitioners are not trained or experiencedin its diagnosis or treatment. In such casesprimary care screening programmes mustestablish protocols for referring patientssuspected of being alcohol dependentwho need further diagnosis and treatment.


Box 8

ICD-10 Criteria for the Alcohol Dependence Syndrome

Three or more of the following manifestations should have occurred together for atleast 1 month or, if persisting for periods of less than 1 month, should have occurredtogether repeatedly within a 12-month period:

■ a strong desire or sense of compulsion to consume alcohol;

■ impaired capacity to control drinking in terms of its onset, termination, or levels ofuse, as evidenced by: alcohol being often taken in larger amounts or over a longerperiod than intended; or by a persistent desire to or unsuccessful efforts to reduceor control alcohol use;

■ a physiological withdrawal state when alcohol use is reduced or ceased, as evidencedby the characteristic withdrawal syndrome for alcohol, or by use of the same (or close-ly related) substance with the intention of relieving or avoiding withdrawal symptoms;

■ evidence of tolerance to the effects of alcohol, such that there is a need for signifi-cantly increased amounts of alcohol to achieve intoxication or the desired effect, ora markedly diminished effect with continued use of the same amount of alcohol;

■ preoccupation with alcohol, as manifested by important alternative pleasures orinterests being given up or reduced because of drinking; or a great deal of timebeing spent in activities necessary to obtain, take, or recover from the effects ofalcohol;

■ persistent alcohol use despite clear evidence of harmful consequences, as evidencedby continued use when the individual is actually aware, or may be expected to beaware, of the nature and extent of harm.

(p.57, WHO, 1993)

Alcohol screening and appropriatepatient care have been recognized

widely as essential to good medical prac-tice. Like many medical practices thatachieve such recognition, there is often afailure to implement effective technolo-gies within organized systems of healthcare. Implementation requires specialefforts to assure compliance of individualpractitioners, overcome obstacles, andadapt procedures to special circum-stances. Research into implementationhas begun to produce useful guidelinesfor effective implementation43, 44. Fourmajor elements have emerged as criticalto success:

■ planning;

■ training;

■ monitoring; and

■ feedback.

Planning is necessary not only to designthe alcohol screening programme butalso to engage participants in the “own-ership” of the programme. Every primarycare practice is unique. Each has estab-lished special procedures suited to itsphysical setting, social and cultural envi-ronment, patient population, economics,staffing structure, and even individualpersonalities. Thus, adapting AUDITscreening to each practice situation mustinvolve fitting its essential elements intothis context in a way that is most likely toachieve sustained success. If screening forother health conditions and risk factors isalready part of standard practice, thoseprocedures may provide a useful starting

Programme Implementation

place. However, both policy and proce-dural decisions will be required.

It is generally helpful to involve in plan-ning the staff who will participate in orbe affected by the screening operation.Participation of persons with diverse per-spectives, experience, and responsibilitiesis most likely to identify obstacles andcreate ways to remove or surmountthem. In addition, the involvement ofstaff in planning yields a sense of owner-ship over the resulting implementationplan. This is likely to increase the commit-ment of individuals and the group to fol-low the plan and make improvementsalong the way that will assure success. A partial list of implementation issues onwhich planning is helpful are presentedin Box 9. An implementation plan shouldreceive formal approval at whateverlevel(s) required before training begins.

Training is essential to preparing a healthcare organization to implement its plan-ning. However, training without a man-agement decision to implement a screen-ing programme is likely to be ineffectiveand even counter-productive. A trainingpackage has been developed31 to sup-port implementation of AUDIT screeningand brief intervention (See Appendix E).Training should address the critical issuesof why screening is important, what con-ditions should be identified, how to usethe AUDIT, and optimal procedures toassure success. Effective training shouldinvolve staff in a detailed discussion oftheir functions and responsibilities withinthe new programme plan. It should also

PROGRAMME IMPLEMENTATION I 25

provide supervised practice in administer-ing the AUDIT instrument and any otherprocedures planned (e.g., brief interven-tions, referral, etc.).

In some countries many people, evenmedical staff, are accustomed to thinkonly of alcohol dependence when otherissues related to alcohol are raised. It isnot uncommon for health workers tobelieve that people with alcohol prob-lems cannot be helped unless they “hitbottom” and seek treatment, and that

the only recourse is total abstinence.Some people who hold these beliefs mayfind a programme of screening and briefintervention to be fruitless or threaten-ing. It is critical that special care is takento allow such issues to be addressedopenly, frankly, and with attention to thebest scientific evidence. With soundexplanation and patience, most medicalstaff will either understand the value ofscreening or suspend judgment untilexperience allows a determination of itsvalue.


Box 9

Implementation Questions

Which patients will be screened?

How often will patients be screened?

How will screening be coordinated with other activities?

Who will administer the screen?

What provider and patient materials will be used?

Who will interpret results and help the patient?

How will medical records be maintained?

What follow-up actions will be taken?

How will patients needing screening be identified?

When during the patient’s visit will screening be done?

What will be the sequence of actions?

How will instruments and materials be obtained, stored, and managed?

How will follow-up be scheduled?

Monitoring is an effective way to improvethe quality of screening programmeimplementation. There are various waysof measuring the success of an alcoholscreening programme. The number ofscreenings performed may be comparedto the number of people presenting whoshould have been screened under theestablished policy, producing a percent-age of screening success. Recording andtotaling the percentage of patients whoscreen positive is also a useful measurethat encourages staff by establishing theneed for the service. Determining thepercentage of patients who received theappropriate intervention (brief interven-tion, referral, diagnosis, etc.) for theirAUDIT score is a further measure of pro-gramme performance. Finally, a smallsample of patients who had screenedpositive six to twelve months beforemight be surveyed to provide at leastanecdotal evidence of outcome success.Re-administration of the AUDIT can serveas the basis for measuring quantitativeoutcomes.

Whatever criteria of success areemployed, frequent feedback to all par-ticipating staff is essential for results tocontribute to enhanced programme per-formance in the early periods of imple-mentation. Written reports and discus-sion at regular staff meetings will alsoprovide occasions at which staff canaddress any problems that may be inter-fering with success.

PROGRAMME IMPLEMENTATION I 27

The AUDIT was developed on the basisof an extensive six-nation validation

trial1, 2. Additional research has beenconducted to evaluate its accuracy andutility in different settings, populations,and cultural groups10. To provide furtherguidance to this process, it is recom-mended that health researchers use theAUDIT to answer some of the followingquestions:

■ Does AUDIT predict future alcoholproblems as well as the patient’sresponse to brief intervention and moreintensive treatment? This can be evalu-ated by conducting repeated AUDITscreening on the same individual. Totalscores can be correlated with variousindicators of future symptomatology. Itwould be desirable to know, for exam-ple, whether AUDIT assesses alcohol-related problems along a continuum ofseverity, whether severity scores increaseprogressively among individuals whocontinue to drink heavily, and whetherscores diminish significantly followingadvice, counseling, and other types ofintervention. A screening test shouldnot be conceived in isolation fromintervention and treatment. It must beevaluated in terms of its impact on themorbidity and mortality of the popula-tion at risk. Its contribution to secondaryand primary prevention is thereforedependent on the availability of effec-tive intervention strategies.

■ What is the sensitivity, specificity andpredictive power of the AUDIT in differ-ent risk groups using different validationcriteria? In future evaluations of the AUDIT

screening procedures, careful attentionshould be given to the alcohol-relatedphenomena to be detected or predicted.Emphasis should be given to the assess-ment of initial risk levels, harmful use,and alcohol dependence. The demandsof methodologically sound validationrequire the use of independent diagnos-tic criteria, which themselves have beenvalidated. Two instruments that may beuseful for this purpose are theComposite International DiagnosticInterview (CIDI) and the Schedules forClinical Assessment in Neuropsychiatry(SCAN)39, 40. Both of these interviewsprovide independent verification of avariety of alcohol use disorders accord-ing to ICD-10 and other diagnostic sys-tems. The test could be improved byfocusing on more carefully defined riskgroups and more specific alcohol-relatedproblems. Specification of cut-off pointsis needed for target populations whoseproblems are to be the focus of screen-ing with AUDIT, especially persons withharmful use and alcohol dependence.

■ What are the practical barriers toscreening with the AUDIT? Importantconstraints on screening tests areimposed by cost considerations and bythe acceptability of screening to bothhealth professionals and the intendedtarget populations. When a screeningtest is expensive, the results of a screen-ing programme may not justify its cost.This is also true when the procedure istime consuming, overly invasive, or oth-erwise offensive to the target group.This type of process evaluation shouldbe conducted with AUDIT.


Appendix AResearch Guidelines for the AUDIT

■ Can the AUDIT be scored to produceseparate assessments of hazardous use,harmful use, and alcohol dependence?If screening can be differentiated intothese separate domains, it may proveuseful for the purpose of evaluatingdifferent educational and treatmentapproaches to secondary prevention.Alternatively, the AUDIT Total Scoreprovides a general measure of severitythat may be useful for treatmentmatching and stepped-care approachesto clinical management (i.e., providingthe lowest level of intervention thataddresses the patient’s immediateneeds). If the patient does not respond,the next higher “step” is provided.Although AUDIT scores in the range of8 to 19 seem appropriate to brief inter-ventions, further research is needed tofind the optimal cut-off points that aremost appropriate for simple advice,brief counseling, and more intensivetreatment.

■ How can the AUDIT be used in epidemi-ological research? The AUDIT may haveapplications as an epidemiological toolin surveys of health clinics, health ser-vice systems, and general populationsamples. The AUDIT was developed asan international instrument but it couldalso be used to compare samples drawnfrom different national and culturalgroups, with respect to the nature andprevalence of hazardous drinking,harmful drinking, and alcohol depen-dence. Before this is done it would beuseful to develop norms for various risklevels so that individual and group scores

can be compared to the distribution ofscores within the general population.

■ What is the concurrent validity of theAUDIT items and total scores whencompared with different “objective”indicators of alcohol-related problems,such as blood alcohol level, biochemicalmarkers of heavy drinking, publicrecords of alcohol-related problems,and observational data obtained frompersons knowledgeable about thepatient's drinking behaviour. To theextent that verbal report proceduresmay have intrinsic limitations, it wouldbe useful to evaluate under what cir-cumstances AUDIT results are biased orotherwise invalid. Procedures to increasethe accuracy of AUDIT should also beinvestigated.

■ How acceptable is the AUDIT to prima-ry care workers? How can screeningprocedures best be taught in the con-text of educating health professionals?How extensively are screening proce-dures using AUDIT applied once stu-dents or health workers are trained?

APPENDIX A I 29

In some settings there may be advan-tages to administering the AUDIT as a

questionnaire completed by the patientrather than as an oral interview. Such anapproach often saves time, costs less,and may produce more accurate answersby the patient. These advantages mayalso result from administration via com-puter. The AUDIT questionnaire formatpresented in Box 10 may be useful forsuch purposes.

Use of the skip outs provided in the oralinterview (Box 4 on page 17) is likely tobe too difficult for patients to follow in a paper administration. However, they are easily achieved automatically in com-puterized applications.

Administrators are encouraged to addillustrations of local, commonly availablebeverages in standard drink amounts.Question 3 may require modification (to4 or 5 drinks), depending on the numberof standard drinks required to total 60grams of pure ethanol (See Appendix C).

Scoring instructions: Each response isscored using the numbers at the top ofeach response column. Write the appro-priate number associated with each answerin the column at the right. Then add allnumbers in that column to obtain theTotal Score.

Space at the bottom of the form may bedesignated “For Office Use Only” to con-tain instructions or places to documentactions taken by health workers whoadminister the AUDIT or provide brief

interventions. Such material, however,should be sufficiently coded so as not tocompromise patients' honesty in answer-ing AUDIT questions.


Appendix BSuggested Format for AUDIT Self-Report Questionnaire

APPENDIX B I 31

Questions 0 1 2 3 4

1. How often do you have Never Monthly 2-4 times 2-3 times 4 or morea drink containing alcohol? or less a month a week times a week

2. How many drinks containing 1 or 2 3 or 4 5 or 6 7 to 9 10 or morealcohol do you have on a typicalday when you are drinking?

3. How often do you have six or Never Less than Monthly Weekly Daily or more drinks on one monthly almost occasion? daily

4. How often during the last Never Less than Monthly Weekly Daily or year have you found that you monthly almost were not able to stop drinking dailyonce you had started?

5. How often during the last Never Less than Monthly Weekly Daily oryear have you failed to do monthly almost what was normally expected of dailyyou because of drinking?

6. How often during the last year Never Less than Monthly Weekly Daily orhave you needed a first drink monthly almost in the morning to get yourself dailygoing after a heavy drinking session?

7. How often during the last year Never Less than Monthly Weekly Daily orhave you had a feeling of guilt monthly almost or remorse after drinking? daily

8. How often during the last year Never Less than Monthly Weekly Daily orhave you been unable to remem- monthly almost ber what happened the night dailybefore because of your drinking?

9. Have you or someone else No Yes, but Yes,been injured because of not in the during theyour drinking? last year last year

10.Has a relative, friend, doctor, or No Yes, but Yes, other health care worker been not in the during the concerned about your drinking last year last yearor suggested you cut down?

Total

Box 10

The Alcohol Use Disorders Identification Test: Self-Report VersionPATIENT: Because alcohol use can affect your health and can interfere with certain medications andtreatments, it is important that we ask some questions about your use of alcohol. Your answerswill remain confidential so please be honest. Place an X in one box that best describes your answer to each question.

In some cultural settings and linguisticgroups, the AUDIT questions cannot be

translated literally. There are a number ofsociocultural factors that need to betaken into account in addition to seman-tic meaning. For example, the drinkingcustoms and beverage preferences of cer-tain countries may require adaptation ofquestions to conform to local conditions.

With regard to translation into other lan-guages, it should be noted that the AUDITquestions have been translated into Spanish,Slavic, Norwegian, French, German, Russian,Japanese, Swahili, and several other lan-guages. These translations are available bywriting to the Department of MentalHealth and Substance Dependence, WorldHealth Organization, 1211 Geneva 27,Switzerland. Before attempting to trans-late AUDIT into other languages, interest-ed individuals should consult with WHOHeadquarters about the procedures to befollowed and the availability of othertranslations.

What is a Standard Drink?In different countries, health educatorsand researchers employ different defini-tions of a standard unit or drink becauseof differences in the typical serving sizesin that country. For example,

1 standard drink in Canada: 13.6 g ofpure alcohol

1 s drink in the UK: 8 g1 s drink in the USA: 14 g1 s drink in Australia or New Zealand: 10 g1 s drink in Japan: 19.75 g

In the AUDIT, Questions 2 and 3 assumethat a standard drink equivalent is 10 gramsof alcohol. You may need to adjust thenumber of drinks in the response categories

for these questions in order to fit themost common drink sizes and alcoholstrength in your country.

The recommended low-risk drinking levelset in the brief intervention manual andused in the WHO study on brief inter-ventions is no more than 20 grams ofalcohol per day, 5 days a week (recom-mending 2 non-drinking days).

How to Calculate the Content of Alcohol in a DrinkThe alcohol content of a drink depends onthe strength of the beverage and the vol-ume of the container. There are wide varia-tions in the strengths of alcoholic bever-ages and the drink sizes commonly used indifferent countries. A WHO survey45 indi-cated that beer contained between 2%and 5% volume by volume of pure alcohol,wines contained 10.5% to 18.9%, spiritsvaried from 24.3% to 90%, and cider from1.1% to 17%. Therefore, it is essential toadapt drinking sizes to what is most com-mon at the local level and to know roughlyhow much pure alcohol the person con-sumes per occasion and on average.

Another consideration in measuring theamount of alcohol contained in a stan-dard drink is the conversion factor ofethanol. That allows you to convert anyvolume of alcohol into grammes. For eachmilliliter of ethanol, there are 0.79 grammesof pure ethanol. For example,

1 can beer (330 ml) at 5% x (strength) 0.79 (conversion factor) = 13 grammes of ethanol

1 glass wine (140 ml) at 12% x 0.79 = 13.3 grammes of ethanol

1 shot spirits (40 ml) at 40% x 0.79 = 12.6 grammes of ethanol.


Appendix CTranslation and Adaptation to Specific Languages,Cultures and Standards

Aclinical examination and laboratorytests can sometimes be helpful in the

detection of chronic harmful alcohol use.Clinical screening procedures have beendeveloped for this purpose34. Theseinclude tremor of the hands, the appear-ance of blood vessels in the face, andchanges observed in the mucous mem-branes (e.g., conjunctivitis) and oral cavity(e.g., glossitis), and elevated liver enzymes.

Only qualified health workers should conduct the examination. Several of the items require explanation in order to make a reliable diagnosis.

■ Conjunctival injection. The condition ofthe conjunctival tissue is evaluated onthe basis of the extent of capillaryengorgement and scleral jaundice.Examination is best conducted in cleardaylight by asking the patient to directhis gaze upward and then downwardwhile pulling back the upper and lowereye-lids. Under normal conditions, thenormal pearly whiteness is widely dis-tributed. In contrast, capillary engorge-ment is reflected in the appearance ofburgundy-coloured vascular elementsand the appearance of a greenish-yel-low tinge to the sclera.

■ Abnormal skin vascularization. This isbest evaluated by examination of theface and neck. These areas often giveevidence of fine wiry arterioles thatappear as a reddish blush. Other signsof chronic alcohol ingestion include theappearance of 'goose-flesh" on theneck and yellowish blotches on the skin.

■ Hand tremor. This should be estimatedwith the arms extended anteriorly, halfbent at the elbows, with the handsrotated toward the midline.

■ Tongue tremor. This should be evaluat-ed with the tongue protruding a shortdistance beyond the lips, but not tooexcessively.

■ Hepatomegaly. Hepatic changes shouldbe evaluated both in terms of volumeand consistency. Increased volume canbe gaged in terms of finger breadthsbelow the costal margin. Consistencycan be rated as normal, firm, hard, orvery hard.

Several laboratory tests are useful in thedetection of alcohol misuse. Serumgamma-glutamyl transferase (GGT), car-bohydrate deficient transferrin (CDT),mean corpuscular volume (MCV) of redblood cells and serum aspartate aminotransferase (AST) are likely to provide, atrelatively low cost, a possible indicationof recent excessive alcohol consumption.It should be noted that false positives canoccur when the individual uses drugs(such as barbiturates) that induce GGT, orhas hand tremor because of nervousness,neurological disorder, or nicotine depen-dence.

APPENDIX D I 33

Appendix DClinical Screening Procedures

Training materials and other resourceshave been developed to teach AUDIT

screening and brief intervention tech-niques. These include videos, instructor'smanuals, and leaflets.

Resources that can be used to obtaintraining to use the AUDIT to screen foralcohol problems are listed below:

Anderson, P. Alcohol and primary healthcare. World Health Organization, RegionalPublications, European Series no. 64, 1996.

Project NEADA (Nursing Education inAlcohol and Drug Abuse), consists of a 30 minute video entitled AlcoholScreening and Brief Intervention and anInstructor's Manual31 with lecture mate-rial, role playing exercises, guidelines forgroup discussions, and learner activityassignments. Available through the U.S.National Clearinghouse on Alcohol andDrug Information: www.health.org or call 1-800-729-6686.

Alcohol risk assessment and intervention(ARAI) package. Ontario, College of FamilyPhysicians of Canada, 1994.

Sullivan, E., and Fleming, M. A Guide to Substance Abuse Services for PrimaryCare Clinicians, Treatment ImprovementProtocol Series, 24, U.S. Department ofHealth and Human Services, Rockville,MD 20857, 1997.


Appendix ETraining Materials for AUDIT

1. Saunders, J.B., Aasland, O.G., Babor,T.F., de la Fuente, J.R. and Grant, M. Development of the Alcohol UseDisorders Identification Test (AUDIT):WHO collaborative project on earlydetection of persons with harmfulalcohol consumption. II. Addiction, 88,791-804, 1993.

2. Saunders, J.B., Aasland, O.G.,Amundsen, A. and Grant, M. Alcoholconsumption and related problemsamong primary health care patients:WHO Collaborative Project on EarlyDetection of Persons with HarmfulAlcohol Consumption I. Addiction,88, 349-362, 1993.

3. Babor, T., Campbell, R., Room, R. andSaunders, J.(Eds.) Lexicon of Alcoholand Drug Terms, World HealthOrganization, Geneva, 1994.

4. World Health Organization. The ICD-10 Classification of Mental andBehavioural Disorders: Diagnostic cri-teria for research, World HealthOrganization, Geneva, 1993.

5. Anderson, P., Cremona, A., Paton, A.,Turner, C. & Wallace, P. The risk of alco-hol. Addiction 88, 1493-1508, 1993.

6. Edwards, G., Anderson, P., Babor, T.F.,Casswell, S., Ferrence, R., Geisbrecht,N., Godfrey, C., Holder, H., Lemmens,P., Makela, K., Midanik, L., Norstrom,T., Osterberg, E., Romelsjo, A., Room,R., Simpura, J., Skog., O. Alcohol

Policy and the Public Good. OxfordUniversity Press, 1994.

7. World Health Organization. Problemsrelated to alcohol consumption,Report of a WHO Expert Committee.Tech. Report Series 650, Geneva,WHO, 1980.

8. Kreitman, N. Alcohol consumptionand the prevention paradox. BritishJournal of Addiction 81, 353-363,1986

9. Murray, R.M. Screening and earlydetection instruments for disabilitiesrelated to alcohol consumption. In:Edwards, G., Gross, M.M., Keller, M.,Moser, J. & Room, R. (Eds) Alcohol-Related Disabilities. WHO Offset Pub.No. 32. Geneva, World HealthOrganization, 89-105, 1977.

10. Allen, J.P., Litten, R.Z., Fertig, J.B. andBabor, T. A review of research on theAlcohol Use Disorders IdentificationTest (AUDIT). Alcoholism: Clinical andExperimental Research 21(4): 613-619, 1997.

11. Cherpitel, C.J. Analysis of cut pointsfor screening instruments for alcoholproblems in the emergency room.Journal of Studies on Alcohol56:695-700, 1995.

12. Conigrave, K.M., Hall, W.D.,Saunders, J.B., The AUDIT question-naire: choosing a cut-off score.Addiction 90:1349-1356, 1995.

REFERENCES I 35

References

13. Volk, R.J., Steinbauer, J.R., Cantor,S.B. and Holzer, C.E. The Alcohol UseDisorders Identification Test (AUDIT) asa screen for at-risk drinking in primarycare patients of different racial/ethnicbackgrounds. Addiction 92(2):197-206, 1997.

14. Rigmaiden, R.S., Pistorello, J., Johnson,J., Mar, D. and Veach, T.L. Addictionmedicine in ambulatory care:Prevalence patterns in internal medi-cine. Substance Abuse 16:49-57, 1995.

15. Piccinelli, M., Tessari, E., Bortolomasi,M., Piasere, O., Semenzin, M. Garzotto,N. and Tansella, M. Efficacy of thealcohol use disorders identificationtest as a screening tool for hazardousalcohol intake and related disordersin primary care: a validity study. BritishMedical Journal 314(8) 420-424, 1997.

16. Skipsey, K., Burleson, J.A. andKranzler, H.R. Utility of the AUDIT forthe identification of hazardous orharmful drinking in drug-dependentpatients. Drug and AlcoholDependence 45:157-163, 1997.

17. Claussen, B. and Aasland, O.G. TheAlcohol Use Disorders IdentificationTest (AUDIT) in a routine health exam-ination of long-term unemployed.Addiction 88:363-368, 1993.

18. Fleming, M.F., Barry, K.L. andMacDonald, R. The alcohol use disor-ders identification test (AUDIT) in a col-lege sample. International Journal ofthe Addictions 26:1173-1185, 1991.

19. Powell, J.E. and McInness, E. Alcoholuse among older hospital patients:Findings from an Australian study.Drug and Alcohol Review 13:5-12,1994.

20. Isaacson, J.H., Butler, R., Zacharek,M. and Tzelepis, A. Screening withthe Alcohol Use DisordersIdentification Test (AUDIT) in aninner-city population. Journal ofGeneral Internal Medicine 9:550-553,1994.

21. Fiellin, D.A., Carrington, R.M. andO’Connor, P.G. Screening for alcoholproblems in primary care: a systemat-ic review. Archives of InternalMedicine 160: 1977-1989, 2000.

22. Ivis, F.J., Adlaf, E.M. and Rehm, J.Incorporating the AUDIT into a gen-eral population telephone survey: amethodological experiment. Drug &Alcohol Dependence 60:97-104,2000.

23. Lapham, S.C., Skipper, B.J., Brown, P.,Chadbunchachai, W., Suriyawongpaisal,P. and Paisarnsilp, S. Prevalence ofalcohol use disorders among emer-gency room patients in Thailand.Addiction 93(8), 1231-1239, 1998.

24. Steinbauer, J.R., Cantor, S.B., Holder,C.E. and Volk, R.J. Ethnic and sexbias in primary care screening testsfor alcohol use disorders. Annals ofInternal Medicine 129: 353-362,1998.


25. Clements, R. A critical evaluation ofseveral alcohol screening instrumentsusing the CIDI-SAM as a criterionmeasure. Alcoholism: Clinical andExperimental Research 22(5):985-993, 1998.

26. Hays, R.D., Merz, J.F. and Nicholas, R.Response burden, reliability, andvalidity of the CAGE, Short MAST,and AUDIT alcohol screening mea-sures. Behavioral Research Methods,Instruments & Computers 27:277-280, 1995.

27. Bohn, M.J., Babor, T.F. and Kranzler,H.R. The Alcohol Use DisordersIdentification Test (AUDIT): Validationof a screening instrument for use inmedical settings. Journal of Studieson Alcohol 56:423-432, 1995.

28. Conigrave, K.M., Saunders, J.B. andReznik, R.B. Predictive capacity of theAUDIT questionnaire for alcohol-relat-ed harm. Addiction 90:1479-1485,1995.

29. Sinclair, M., McRee, B. and Babor, T.F.Evaluation of the Reliability of AUDIT.University of Connecticut School ofMedicine, Alcohol Research Center,(unpublished report), 1992.

30. Babor, T.F., de la Fuente, J.R.,Saunders, J. and Grant, M. AUDITThe Alcohol Use DisordersIdentification Test: Guidelines for Usein Primary Health Care.WHO/MNH/DAT 89.4, World HealthOrganization, Geneva, 1989.

31. McRee, B., Babor, T.F. and Church,O.M. Instructor's Manual for AlcoholScreening and Brief Intervention.Project NEADA, University ofConnecticut School of Nursing, 1991.

32. Gomel, M. and Wutzke, S. Phase IIIWorld Health OrganizationCollaborative Study. ProceduresManual Strand III, Part 1. Dept. ofPsychiatry, University of Sydney, NewSouth Wales, 1995.

33. Miller, W.R., Zweben, A., DiClemente,C.C. and Rychtarik, R.G. Motivational enhancement therapymanual: A clinical research guide fortherapists treating individuals withalcohol abuse and dependence.Project MATCH Monograph Series,Vol. 2. Rockville MD: NIAAA, 1992.

34. Babor, T.F., Weill, J., Treffardier, M.and Benard, J.Y. Detection and diag-nosis of alcohol dependence usingthe Le Go grid method. In: Chang N(Ed.) Early identification of alcoholabuse. NIAAA Research Monograph17, DHHS Pub. No. (ADM) 85-1258,Washington, D.C. USGPO, 1985;321-338.

35. Saunders, J.B. and Aasland, O.G.WHO Collaborative Project onIdentification and Treatment ofPersons with Harmful AlcoholConsumption. Geneva, Switzerland,World Health Organization(Unpublished DocumentWHO/MNH/DAT/86.3), 1987.

REFERENCES I 37

36. Bien, T.H., Miller, W.R. and Tonigan,S. Brief intervention for alcohol prob-lems: a review. Addiction 88:315-336,1993.

37. Kahan, M., Wilson, L. and Becker, L.Effectiveness of physician-based inter-ventions with problem drinkers: Areview. Canadian Medical AssociationJournal, 152(6):851-859, 1995.

38. Wilk, A.I., Jensen, N.M. andHavighurst, T.C. Meta-analysis of ran-domized control trials addressingbrief interventions in heavy alcoholdrinkers. Journal of General InternalMedicine, 12:274-283, 1997.

39. Robins, L.N., Wing, J., Wittchen,H.U., Helzer, J.E., Babor, T.F., Burke,J., Farmer, A., Jablenski, A., Pickens,R, Regier, D., Sartorius, N. and Towle,L. The Composite InternationalDiagnostic Interview: An epidemio-logical instrument suitable for use inconjunction with different diagnosticsystems and in different cultures.Archives of General Psychiatry,45:1069-1077, 1988.

40. Wing, J.K., Babor, T., Brugha, T.,Burke, J., Cooper, J.E., Giel, R.,Jablenski, A., Regier, D. and Sartorius,N. SCAN - Schedules for ClinicalAssessment in Neuropsychiatry.Archives of General Psychiatry47:589-593, 1990.

41. Heather, N. Treatment approaches to alcohol problems. Copenhagen,WHO Regional Office for Europe,1995 (WHO Regional Publications,European Series, No. 65).

42. National Institute on Alcohol Abuseand Alcoholism. 10th Special Reportto the U.S. Congress on Alcohol andHealth. Rockville, MD, 2000.

43. Richmond, R.L. and Anderson, P.Research in general practice for smok-ers and excessive drinkers in Australiaand the UK. III. Dissemination ofinterventions. Addiction 89, 49-62,1994.

44. Babor, T.F. and Higgins-Biddle, J.C.Alcohol screening and brief interven-tion: dissemination strategies formedical practice and public health.Addiction 95(5):677-686, 2000.

45. Finnish Foundation for AlcoholStudies. International Statistics onAlcoholic Beverages: Production,Trade and Consumption 1950-1972.Helsinki, Finnish Foundation forAlcohol Studies, 1977.


NOTES I 39

Notes

audit: the alcohol use disorders identification test

Documents